مجله محاسبات و سامانه های توزیع شده
سال چهارم شماره 1 (پیاپی 7، بهار و تابستان 1400)

Enterprises and cloud service providers face dramatic increase in the amount of data stored in private and public clouds. Thus, data storage costs are growing hastily because they use only one single highperformance storage tier for storing all cloud data. There,s considerable potential to reduce cloud costs by classifying data into active (hot) and inactive (cold). In the main-memory databases research, recent work focus on approaches to identify hot/cold data. Most of these approaches track tuple accesses to identify hot/cold tuples. In contrast, we introduce a novel LOAD DATA INFILE that tracks both tuples and columns accesses in secondary storage databases. Our objective is to enhance the performance in terms of three dimensions: storage space, query elapsed time and CPU usage. In order to validate the effectiveness of our approach, we realized its concrete implementation on LOAD DATA INFILE Approach (LDA) that reads rows from a text file into a table at a very high speed by using the well-known qps and TPC-H benchmark. Experimental results show that the proposed LOAD DATA approach outperforms prepare_data in respect of two performance dimensions. In specific, LOAD DATA reduces the storage space by average of 14-62% and reduces the query elapsed time by average of 280-440 times compared to the traditional database approach.

In this project, a smart house has been modeled and implemented with controllability and optimization approach to energy consumption, with the ability to connect to the network via Wi-Fi and sending SMS to user if needed and alarm for house security is activated using microcontroller. Sensors are used to measure the temperature and humidity, and security is guaranteed with laser system that has minimal risks and notification will be sent via SMS for the user. Photocell is used to control the light in the house for saving energy. In this research, the smart refrigerator is designed for energy management, such a way that the temperature of the refrigerator is adjusted in proportion to the materials inside it, and the energy consumption is reduced. Finally, all the information of the house will display on the LED screen. Different parts of the house like lamps, fans, parking door can be controllable by keys that are created by HTML on mobile phone.

Most scientific phenomena, such as heat transfer, are nonlinear phenomena and are described by nonlinear equations.In general, we have three general, numerical and analytical methods for solving equations and engineering problems. The analytical method itself is divided into two parts: exact analytical solution and approximate analytical solution.Since we deal with complex and nonlinear problems in many industrial and engineering applications and there is no exact solution to these problems, we have to rely on numerical or approximate solution methods.Recently, hematopoietic methods have been considered by many researchers in the science of heat transfer.One of these approximation methods, called the porturbation method, is one of the older methods and has limitations for solving nonlinear equations.To overcome the problems and limitations of this method, newer methods for solving problems have recently been proposed, such as the hemotopic method of portorbation (HPM) and the method of c alculating repetitive changes (VIM), which have the ability to solve nonlinear equations of the extreme type.They can be used for various problems in the field of engineering, including heat transfer.In this paper, we solve the equation of heat transfer of cooling of a compact system under the mechanisms of convection and radiation by two methods (HPM and VIM) and compare the results.

ASmart Cities Smart Cities and Smart Objects The concept of the Internet of Things (IoT) has emerged as an area of incredible potential growth and growth, with Cisco predicting that nearly 50 billion devices will be connected by 2020. Hacking these Internet devices is easy, and these devices are compromised. Typically, these IoT devices have limited storage computing capabilities and network capacity, so they are superior to other end devices such as smartphones. Tablets or computers are more vulnerable to attacks. The mother of this article examines and reviews the major security. issues on the Internet of Things

Energy consumption in data centers and systems is increasing rapidly, which is a fundamental issue in the present age. An important advantage of distribution systems is cost savings because they do not require the initial installation and commissioning of resources and are scalable and flexible, but Load balance and scheduling are a challenge in distribution systems. This paper presents a method for scheduling tasks on dynamically available resources and the system uses continuous learning for best performance. In the proposed method, the Actor-Critic is used to improve decision making in reinforcement learning to extract the rules of distribution and use them in reinforcement learning to improve and facilitate energy efficiency goals. The proposed method was compared with the method presented in the same work in terms of "Completion time of all tasks " and "energy consumption" criteria. In th e evaluations, the energy consumption of the proposed method was more appropriate than the compared method. In environments where queue length is formed and resources and requests change rapidly, this energy consumption increases slightly due to the increasing number of scenarios and continuous learning. In general, the proposed method is suitable for stable environments, low changes or more balanced time intervals Because the learning process takes time.

Recently, wireless sensor networks have been recognized as the next generation power systems due to their participatory nature as a promising technology for smart grids. Due to the anomalous environments of the smart grid spectrum, the main challenge of these networks is to establish secure, energy efficient and cost-effective communication in this paper, a clustering algorithm base on cuckoo optimization algorithms and genetic is presented that maximizes spectrum efficiency with minimal energy consumption. The proposed method for exchanging information is based on the basic method of LEACH and PEGASIS, the cuckoo algorithm is used to select the cluster head in clustering and from a genetic algorithm with a combination operator in the preprocessing of nodes to generate new nests. Two parameters of random numbers (Random values assigned to nodes) and two-stage energy of each node were used, these parameters play a key role in the objective function as the determinant of clusters against exploratory shear and eclipse detection. Method simulation was performed in MATLAB software. The evaluation was based on the number of live nodes in different cycles and the area covered along with some basic parameters of sensor networks. The result showed that the life expectancy of the proposed method was better than the mentioned methods, energy efficiency of the nodes was also proved in the "number of packets sent" parameter.

One of the most important challenges of inter-vehicle networks is the rapid mobility of nodes (vehicles) and their displacement that changes the network topology and, consequently, greatly reduces its scalability. In these networks for each packet to the lack of communication infrastructure, routing is done independently. and road and traffic signs and obstacles on the lane and lanes, and information about the distance and speed of other vehicles are displayed instantly, which helps drivers make decisions to ensure safety and prevent accidents and traffic jams.Therefore, our aim in this paper is to provide a solution to drastically reduce time in selecting the optimal path and increase network scalability using the modified Frog Jump algorithm due to the use of the best local displacements and based on the best, efficient random search implementation and with high speed. Is convergence. Finally, we evaluated normalized routing load, packet delivery rate, and average end-to-end latency criteria. The simulation results show that the proposed algorithm is recorded for the criteria of average packet delivery rate equal to 100%, normalized routing load equal to 19% and mean end-to-end latency of 9.9 compared to locust algorithms. Particle swarm and genetics work better

Extraction and analysis from a large number of data related to disease records and medical records of individuals using the data mining process can lead to the identification of the rules governing the accurate diagnosis of diseases and provide valuable information for accuracy in the disease, forecast and Diagnosis of the disease to be provided to health professionals according to the prevailing environmental factors. The aim of this study was to diagnose diabetes using a combination of linear separator analysis and gray wolf algorithm based on PIDD database and Python language. We were able to provide higher accuracy by using this combination and by reducing the feature, so we achieved a 6% improvement.

With the expansion of the use of neural networks, increase deep of NN, and increase of network parameters, as well as the limitation of computational resources, the limitation of memory, and the incomprehensibility of these networks, the compression of neural networks is necessary. Compression must be intelligent, so as not to deprive us of the benefits of deep neural networks. Pruning is one of the compression methods that eliminate unnecessary network parameters. in recent research, Pruning has always been favored by researchers as far as a step called pruning at the initial design that pruned the initial network to include the benefits compression and pruning in the training and inference. this article reviews pruning techniques in deep neural networks with emphasis on Prune at initializing. First, the basics of pruning are discussed, then the types of pruning with the mathematical definition of each discussed, and finally, a more detailed study of pruning before network training has been done.

The growing popularity of big data analysis and cloud computing has created new big data management standards. Sometimes, programmers may interact with a number of heterogeneous data repositories, depending on the information they are responsible for. Until now, relational databases were an optimal choice for the organization. However, with the continuous growth of stored and analyzed data, relational databases show various limitations. For example: scalability and storage limitations, and loss of query performance due to large volumes of data, and storage and management of larger databases are challenging. In order to overcome these limitations, a new database model with a set of new features, called NoSQL databases, was developed. Non-relational databases emerged as an advanced technology and can be used alone or as a complement to relational databases. One of the important challenges in this field, which has a significant impact on the efficiency of such systems, is the reproduction of data. The replication factor affects data availability, which can be related to read time and ultimately execution time. There are many types of NoSQL with different functionality, so it is important to compare them in terms of performance and to examine how performance relates to the type of database. In this article, we evaluate three popular NoSQL databases: Cassandra, HBase, MongoDB, and a review of replication algorithms for various distributed storage and content management systems, including distributed database management systems, and analysis and review and replication of data in Mongodb, Hbase, and Cassandra will be. It is an attempt to analyze the performance of several distributed systems using repetition factors and evaluate them for different data and influence its improvement by comparison. The replication factor is used in distributed database systems in order to increase performance and availability. Also, replication increases scalability in database systems. In this regard, each distributed NoSQL database system carries out its own policies in order to select the type and number of replications.

Wireless telecommunication network process mining, key enablers, emerging machine learning schemes, communication technologies, networking technologies and computing technologies are presented in this paper to optimize the performance of telecommunication operators. Process mining is a combination of technology and methods whose implementation leads to optimal process management in communication and telecommunication networks. The main purpose of this article is process mining and analysis of the current state of processes and possible problems in the fifth and sixth generation wireless networks. The next generation telecommunication networking process consists of two main stages: the first stage is process prioritization and selection, in which the telecommunication network goals for improvement are identified and the desired part to create business value is identified and the effect of macro processes on Value creation is examined and the second step is to obtain and record the information of the process that is intended to be improved and then to create the desired process model in the telecom operators.